مرکز منطقه ای اطلاع رساني علوم و فناوري - Selection of anchor pixels by a using image filtering approach for internal calibration step of evapotranspiration mapping through energy balance

Abstract :

Evapotranspiration (ET) has a great importance for calculation of hydrologic water budget in waterresource management and estimation of irrigation water amount in agricultural water management. Inrecent years, ET mapping opportunities have been gradually increased by the development of satellitetechnologies. Almost all models (METRIC, SEBAL etc.) developed for ET mostly depends on energybalance, the main components of which are latent heat flux (LE), sensible heat flux (H), net radiation(Rn) and soil heat flux (G). METRIC and SEBAL are two important energy balance based models. Inorder to run these models, the data obtained from satellite images and meteorological stations should beused together, and one of the most important step is the calculation of sensible heat flux (H). InMETRIC and SEBAL models, the calculation of H contains an internal calibration process whichdepends on two selected anchor pixels called hot and cold. Cold pixel represents a condition of ET atpotential level and hot pixel represents a situation with ET at zero level. Therefore, the selection of hotand cold pixels have an important effect on the accuracy of ET maps. Anchor pixels should be selectedfrom agricultural areas, and the filtration equations are required in the case of the large study area. The main purpose of this study was to develop filtration equations for facilitating the selection of anchor pixels for a current model approach, METRIC and thus improve the accuracy of ET maps. This study was carried out for two separate regions; one of them having sub-humid (Bafra) and the other having semi-arid (Suluova) climate conditions. Landsat 8 satellite images were used for calculations. The equations developed for filtration purposes were based on the normalized difference vegetation index (NDVI) and surface temperature (Ts). By using the filtration method developed in this study, the selection of anchor pixels was achieved correctly and easily. This filtration method standardized the anchor pixel selection, which could vary from one person to another. The application of filtration improved the accuracy of ET maps and offered a significant opportunity to save time.